In [1]:
# http://127.0.0.1:8888/?token=d8645dfa36d67d3f2ca03610ce0c3c8e01cc36bd6dfab090
import utils
import numpy as np
import pandas as pd
pd.set_option('display.max_rows', None)
pd.set_option('display.max_columns', None)
import matplotlib.pyplot as plt
% matplotlib inline
import seaborn as sns
sns.set(style="whitegrid", color_codes=True)
sns.set(font_scale=1.4)
color = sns.color_palette()
from sklearn.decomposition import PCA
读取数据以及数据预览¶
In [2]:
train_data = pd.read_csv(utils.file_train_data)
test_data = pd.read_csv(utils.file_test_data)
print('Shape train_data: {}\nShape test_data: {}'.format(train_data.shape,test_data.shape))
In [ ]:
# train_data.head()
In [ ]:
# test_data.head()
In [3]:
train_data.describe().T
Out[3]:
In [4]:
test_data.describe().T
Out[4]:
数据预处理 train_data to X_train test_data to X_test¶
- 去掉无用列
In [5]:
X_train = train_data.drop(['id','weight','era'], axis=1)
y_train = X_train.pop('label')
X_train_group = train_data['group']
X_test = test_data.drop(['id'], axis=1)
X_test_group = test_data['group']
- One hot encoding
In [ ]:
# group_train_dummies = pd.get_dummies(X_train['group'],prefix='group')
# group_test_dummies = pd.get_dummies(X_test['group'],prefix='group')
In [ ]:
# X_train.drop(['group'],axis=1,inplace=True)
# X_test.drop(['group'],axis=1,inplace=True)
# X_train=X_train.join(group_train_dummies)
# X_test=X_test.join(group_test_dummies)
label列和group列以及weight列统计¶
In [6]:
train_data['label'].value_counts()
Out[6]:
In [7]:
print('label 1 ratio: {}\n'.format(sum(train_data['label']) / train_data.shape[0]))
print('label 0 ratio: {}\n'.format(len(train_data[train_data['label'] == 0]) / train_data.shape[0]))
In [8]:
sns.set(font_scale=1.8)
int_level = train_data['label'].value_counts()
plt.figure(figsize=(20,5))
sns.barplot(int_level.index, int_level.values, alpha=0.8, color='b')
plt.title('Counts of label in train_data')
plt.show()
In [9]:
train_data['group'].value_counts()
Out[9]:
In [10]:
test_data['group'].value_counts()
Out[10]:
In [11]:
sns.set(font_scale=1.8)
int_level = train_data['group'].value_counts()
plt.figure(figsize=(20,5))
sns.barplot(int_level.index, int_level.values, alpha=0.8, color='b')
plt.title('Counts of group in train_data')
plt.show()
In [12]:
sns.set(font_scale=1.8)
int_level = test_data['group'].value_counts()
plt.figure(figsize=(20,5))
sns.barplot(int_level.index, int_level.values, alpha=0.8, color='r')
plt.title('Counts of group in test_data')
plt.show()
In [13]:
train_data['weight'].value_counts()
Out[13]:
In [14]:
sns.set(font_scale=2.2)
int_level = train_data['weight'].value_counts()
plt.figure(figsize=(50,20))
sns.barplot(int_level.index, int_level.values, alpha=0.8, color='b')
plt.title('Counts of weight in train_data')
plt.show()
Feature Correlation 以及 Importance¶
In [15]:
train_data = train_data.drop(['id','weight','era'], axis=1)
In [16]:
corrmat = train_data.corr()
#f, ax = plt.subplots(figsize=(500, 500))
#sns.heatmap(corrmat, vmax=.8, square=True);
In [17]:
k = 15 #number of variables for heatmap
plt.figure(figsize=(20, 15))
cols = corrmat.nlargest(k, 'label')['label'].index
cm = np.corrcoef(train_data[cols].values.T)
sns.set(font_scale=1.75)
hm = sns.heatmap(cm, cbar=True, annot=True, square=True,
fmt='.4f', annot_kws={'size': 15}, yticklabels=cols.values,
xticklabels=cols.values)
plt.show()
In [18]:
numeric_features = train_data.select_dtypes(include=[np.number])
numeric_features.dtypes
corr = numeric_features.corr()
print("%d features in total" %(len(corr)))
print (corr['label'].sort_values(ascending=False)[:90], '\n')
#print (corr['label'].sort_values(ascending=False)[-5:])
查看Feature分布¶
In [46]:
sns.set(font_scale=1.8)
plt.figure(figsize=(20, 10))
sns.distplot(X_train['feature0'], bins=100 ,kde_kws={'lw': 2.5, 'color': 'k', 'alpha': 0.7, 'label': 'KDE'}, hist_kws={ 'histtype': 'step', 'lw': 3, 'color': 'b', 'alpha': 0.9, 'label': 'feature0'})
plt.title('Distribution of Feature0 in X_train')
plt.show()
In [47]:
sns.set(font_scale=1.8)
plt.figure(figsize=(20, 10))
sns.distplot(X_test['feature0'], bins=100 ,kde_kws={'lw': 2.5, 'color': 'k', 'alpha': 0.7, 'label': 'KDE'}, hist_kws={ 'histtype': 'step', 'lw': 3, 'color': 'r', 'alpha': 0.9, 'label': 'feature0'})
plt.show()
In [49]:
sns.set(font_scale=1.8)
plt.figure(figsize=(20, 10))
ax1 = sns.distplot(X_train['feature0'], bins=100, kde=True, color='b', kde_kws={'lw': 3, 'label':'X_train'}, hist=False)
ax2 = sns.distplot(X_test['feature0'], bins=100, kde=True, color='r', kde_kws={'lw': 3, 'label':'X_test'}, hist=False)
plt.title('Distribution of Feature0 in X_train and X_test');
ax1.set_xlim([-2,5])
ax2.set_xlim([-2,5])
plt.show()
In [22]:
sns.set(font_scale=1.8)
plt.figure(figsize=(20, 10))
plt.scatter(range(X_train.shape[0]), train_data["feature0"].values, color='m')
plt.title('Sactter distribution of Feature0 in X_train')
plt.show()
In [23]:
sns.set(font_scale=1.8)
plt.figure(figsize=(20, 10))
plt.scatter(range(X_test.shape[0]), test_data["feature0"].values, color='g')
plt.title('Sactter distribution of Feature0 in X_test')
plt.show()
- 训练集和测试集feature0-87以及group分布对比
In [50]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2, figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[0], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[0], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [51]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[1], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[1], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [52]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[2], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[2], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [53]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[3], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[3], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [54]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[4], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[4], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [55]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[5], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[5], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [56]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[6], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[6], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [57]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[7], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[7], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [58]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[8], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[8], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [59]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[9], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[9], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [60]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[10], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[10], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [61]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[11], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[11], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [62]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[12], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[12], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [63]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[13], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[13], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [64]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[14], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[14], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [65]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[15], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[15], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [66]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[16], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[16], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [67]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[17], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[17], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [68]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[18], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[18], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [69]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[19], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[19], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [70]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[20], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[20], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [71]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[21], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[21], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [72]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[22], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[22], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [73]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[23], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[23], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [74]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[24], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[24], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [75]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[25], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[25], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [76]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[26], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[26], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [77]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[27], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[27], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [78]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[28], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[28], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [79]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[29], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[29], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [80]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[30], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[30], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [81]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[31], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[31], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [82]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[32], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[32], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [83]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[33], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[33], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [84]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[34], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[34], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [85]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[35], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[35], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [86]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[36], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[36], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [87]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[37], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[37], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [88]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[38], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[38], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [89]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[39], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[39], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [90]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[40], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[40], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [91]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[41], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[41], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [92]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[42], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[42], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [93]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[43], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[43], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [94]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[44], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[44], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [95]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[45], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[45], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [96]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[46], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[46], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [97]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[47], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[47], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [98]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[48], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[48], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [99]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[49], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[49], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [100]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[50], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[50], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [101]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[51], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[51], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [102]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[52], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[52], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [103]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[53], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[53], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [104]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[54], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[54], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [105]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[55], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[55], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [106]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[56], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[56], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [107]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[57], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[57], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [108]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[58], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[58], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [109]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[59], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[59], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [110]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[60], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[60], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [111]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[61], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[61], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [112]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[62], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[62], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [113]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[63], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[63], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [114]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[64], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[64], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [97]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[65], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[65], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
Out[97]:
In [115]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[66], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[66], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [116]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[67], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[67], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [117]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[68], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[68], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [118]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[69], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[69], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [119]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[70], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[70], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [120]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[71], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[71], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [121]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[72], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[72], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [122]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[73], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[73], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [123]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[74], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[74], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [124]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[75], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[75], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [125]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[76], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[76], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [126]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[77], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[77], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [127]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[78], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[78], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [128]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[79], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[79], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [129]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[80], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[80], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [130]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[81], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[81], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [131]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[82], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[82], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [132]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[83], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[83], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [133]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[84], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[84], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [134]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[85], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[85], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [135]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[86], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[86], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [136]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train.columns[87], data=X_train, ax=axes[0])
cx = sns.violinplot(y=X_test.columns[87], data=X_test, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
plt.show()
In [137]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=train_data['group'], data=X_train, ax=axes[0])
cx = sns.violinplot(y=test_data['group'], data=X_test, ax=axes[1])
bx.set_ylim([-1,28])
cx.set_ylim([-1,28])
plt.show()
按group列分割表¶
- group=1
In [ ]:
X_train_1 = X_train[X_train['group'] == 1]
X_test_1 = X_test[X_test['group'] == 1]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_1.columns[0], data=X_train_1, ax=axes[0])
cx = sns.violinplot(y=X_test_1.columns[0], data=X_test_1, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_1.columns[1], data=X_train_1, ax=axes[0])
cx = sns.violinplot(y=X_test_1.columns[1], data=X_test_1, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_1.columns[2], data=X_train_1, ax=axes[0])
cx = sns.violinplot(y=X_test_1.columns[2], data=X_test_1, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_1.columns[3], data=X_train_1, ax=axes[0])
cx = sns.violinplot(y=X_test_1.columns[3], data=X_test_1, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=2
In [ ]:
X_train_2 = X_train[X_train['group'] == 2]
X_test_2 = X_test[X_test['group'] == 2]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_2.columns[3], data=X_train_2, ax=axes[0])
cx = sns.violinplot(y=X_test_2.columns[3], data=X_test_2, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=3
In [ ]:
X_train_3 = X_train[X_train['group'] == 3]
X_test_3 = X_test[X_test['group'] == 3]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_3.columns[3], data=X_train_3, ax=axes[0])
cx = sns.violinplot(y=X_test_3.columns[3], data=X_test_3, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=4
In [ ]:
X_train_4 = X_train[X_train['group'] == 4]
X_test_4 = X_test[X_test['group'] == 4]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_4.columns[3], data=X_train_4, ax=axes[0])
cx = sns.violinplot(y=X_test_4.columns[3], data=X_test_4, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=5
In [ ]:
X_train_5 = X_train[X_train['group'] == 5]
X_test_5 = X_test[X_test['group'] == 5]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_5.columns[3], data=X_train_5, ax=axes[0])
cx = sns.violinplot(y=X_test_5.columns[3], data=X_test_5, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=6
In [ ]:
X_train_6 = X_train[X_train['group'] == 6]
X_test_6 = X_test[X_test['group'] == 6]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_6.columns[3], data=X_train_6, ax=axes[0])
cx = sns.violinplot(y=X_test_6.columns[3], data=X_test_6, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=7
In [ ]:
X_train_7 = X_train[X_train['group'] == 7]
X_test_7 = X_test[X_test['group'] == 7]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_7.columns[3], data=X_train_7, ax=axes[0])
cx = sns.violinplot(y=X_test_7.columns[3], data=X_test_7, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=8
In [ ]:
X_train_8 = X_train[X_train['group'] == 8]
X_test_8 = X_test[X_test['group'] == 8]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_8.columns[3], data=X_train_8, ax=axes[0])
cx = sns.violinplot(y=X_test_8.columns[3], data=X_test_8, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=9
In [ ]:
X_train_9 = X_train[X_train['group'] == 9]
X_test_9 = X_test[X_test['group'] == 9]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_9.columns[3], data=X_train_9, ax=axes[0])
cx = sns.violinplot(y=X_test_9.columns[3], data=X_test_9, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=10
In [ ]:
X_train_10 = X_train[X_train['group'] == 10]
X_test_10 = X_test[X_test['group'] == 10]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_10.columns[3], data=X_train_10, ax=axes[0])
cx = sns.violinplot(y=X_test_10.columns[3], data=X_test_10, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=11
In [ ]:
X_train_11 = X_train[X_train['group'] == 11]
X_test_11 = X_test[X_test['group'] == 11]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_11.columns[3], data=X_train_11, ax=axes[0])
cx = sns.violinplot(y=X_test_11.columns[3], data=X_test_11, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=12
In [ ]:
X_train_12 = X_train[X_train['group'] == 12]
X_test_12 = X_test[X_test['group'] == 12]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_12.columns[3], data=X_train_12, ax=axes[0])
cx = sns.violinplot(y=X_test_12.columns[3], data=X_test_12, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=13
In [ ]:
X_train_13 = X_train[X_train['group'] == 13]
X_test_13 = X_test[X_test['group'] == 13]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_13.columns[3], data=X_train_13, ax=axes[0])
cx = sns.violinplot(y=X_test_13.columns[3], data=X_test_13, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=14
In [ ]:
X_train_14 = X_train[X_train['group'] == 14]
X_test_14 = X_test[X_test['group'] == 14]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_14.columns[3], data=X_train_14, ax=axes[0])
cx = sns.violinplot(y=X_test_14.columns[3], data=X_test_14, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=15
In [ ]:
X_train_15 = X_train[X_train['group'] == 15]
X_test_15 = X_test[X_test['group'] == 15]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_15.columns[3], data=X_train_15, ax=axes[0])
cx = sns.violinplot(y=X_test_15.columns[3], data=X_test_15, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=16
In [ ]:
X_train_16 = X_train[X_train['group'] == 16]
X_test_16 = X_test[X_test['group'] == 16]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_16.columns[3], data=X_train_16, ax=axes[0])
cx = sns.violinplot(y=X_test_16.columns[3], data=X_test_16, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=17
In [ ]:
X_train_17 = X_train[X_train['group'] == 17]
X_test_17 = X_test[X_test['group'] == 17]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_17.columns[3], data=X_train_17, ax=axes[0])
cx = sns.violinplot(y=X_test_17.columns[3], data=X_test_17, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=18
In [ ]:
X_train_18 = X_train[X_train['group'] == 18]
X_test_18 = X_test[X_test['group'] == 18]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_18.columns[3], data=X_train_18, ax=axes[0])
cx = sns.violinplot(y=X_test_18.columns[3], data=X_test_18, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=19
In [ ]:
X_train_19 = X_train[X_train['group'] == 19]
X_test_19 = X_test[X_test['group'] == 19]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_19.columns[3], data=X_train_19, ax=axes[0])
cx = sns.violinplot(y=X_test_19.columns[3], data=X_test_19, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=20
In [ ]:
X_train_20 = X_train[X_train['group'] == 20]
X_test_20 = X_test[X_test['group'] == 20]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_20.columns[3], data=X_train_20, ax=axes[0])
cx = sns.violinplot(y=X_test_20.columns[3], data=X_test_20, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=21
In [ ]:
X_train_21 = X_train[X_train['group'] == 21]
X_test_21 = X_test[X_test['group'] == 21]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_21.columns[3], data=X_train_21, ax=axes[0])
cx = sns.violinplot(y=X_test_21.columns[3], data=X_test_21, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=22
In [ ]:
X_train_22 = X_train[X_train['group'] == 22]
X_test_22 = X_test[X_test['group'] == 22]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_22.columns[3], data=X_train_22, ax=axes[0])
cx = sns.violinplot(y=X_test_22.columns[3], data=X_test_22, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=23
In [ ]:
X_train_23 = X_train[X_train['group'] == 23]
X_test_23 = X_test[X_test['group'] == 23]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_23.columns[3], data=X_train_23, ax=axes[0])
cx = sns.violinplot(y=X_test_23.columns[3], data=X_test_23, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=24
In [ ]:
X_train_24 = X_train[X_train['group'] == 24]
X_test_24 = X_test[X_test['group'] == 24]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_24.columns[3], data=X_train_24, ax=axes[0])
cx = sns.violinplot(y=X_test_24.columns[3], data=X_test_24, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=25
In [ ]:
X_train_25 = X_train[X_train['group'] == 25]
X_test_25 = X_test[X_test['group'] == 25]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_25.columns[3], data=X_train_25, ax=axes[0])
cx = sns.violinplot(y=X_test_25.columns[3], data=X_test_25, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=26
In [ ]:
X_train_26 = X_train[X_train['group'] == 26]
X_test_26 = X_test[X_test['group'] == 26]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_26.columns[3], data=X_train_26, ax=axes[0])
cx = sns.violinplot(y=X_test_26.columns[3], data=X_test_26, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=27
In [ ]:
X_train_27 = X_train[X_train['group'] == 27]
X_test_27 = X_test[X_test['group'] == 27]
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_27.columns[3], data=X_train_27, ax=axes[0])
cx = sns.violinplot(y=X_test_27.columns[3], data=X_test_27, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
- group=28
In [ ]:
sns.set(font_scale=1.4)
fg,axes = plt.subplots(1,2,figsize=(20, 5),sharex=True)
bx = sns.violinplot(y=X_train_26.columns[3], data=X_train_26, ax=axes[0])
cx = sns.violinplot(y=X_test_26.columns[3], data=X_test_26, ax=axes[1])
bx.set_ylim([-3,3])
cx.set_ylim([-3,3])
train_data.info()¶
In [ ]:
#train_data.isnull().sum()
In [ ]:
#table_type = train_data.dtypes.reset_index()
#table_type.columns=['feat','type']
#print(table_type.groupby('type').aggregate('count'))
In [ ]:
#sns.countplot(train_data.label,order=['0', '1']);
#plt.xlabel('label');
#plt.ylabel('Counts');
PCA处理¶
- 对[group]列28个虚拟变量做PCA处理
In [ ]:
pca = PCA(n_components=28)
In [ ]:
pca.fit(group_train_dummie)
In [ ]:
print(pca.explained_variance_ratio_)
print(pca.explained_variance_)
print(pca.n_components_)
- 对feature0-feature88做PCA处理
In [ ]:
pca = PCA(n_components=88)
In [ ]:
pca.fit_transform(X_train)
In [ ]:
print(pca.explained_variance_ratio_)
print(pca.explained_variance_)
print(pca.n_components_)
In [ ]:
pca = PCA(n_components=88)